I prefer third option: define two small functions inside the large one. That way, the "root" level of the large function is still small, but you can clearly see

I prefer third option: define two small functions inside the large one. That way, the "root" level of the large function is still small, but you can clearly see the scope in which the two smaller ones are used.